One well known and long-lasting problem in the video tracking is that one particular algorithm would perform well on a certain environmental characteristic. Whenever the characteristic in the scene changes; the performance of the algorithm affected. This research proposes a multiagent-based for video tracking system. The agents follow the odd-man out strategy; which odd agents will be credited less than the favorite ones. We tested our algorithm against two tough videos. The results show that our approach yield satisfactory outcomes. The final tracking results are always within the boundary of the groundtruth; given that there are two out of five correct results.

This paper proposes a solution for network monitoring and forecasting of user behavior activities in social network applications. This method can describe the log of various types of data; behaviors and traces in the power monitoring system. The monitoring system analyzes by using Pentaho BI open source software to extract the data. In this study; the monitoring system analyzes and identifies types of TCP packets. The monitoring and packet capturing system was implemented on the campus's wire and wireless LAN network at the Faculty of Information Technology; King Mongkut's University of Technology North Bangkok during active hours. The system implementation uncovered the percentage of several social media types mostly used in the network during each time period. The result showed that several kinds of data packet such as packet loss; TCP or SYN flooding provided useful information for the network administrator to improve and manage the system.

Shapley value is regarded as a fair payoff distribution concept for cooperative agents. While traditional cooperative game assume superadditivity and non-externalty; real world environments do not hold this assumption. We show that in linear production game; the environment is non-superadditive is with externalties. In such environment; grand coalition does not provide optimal solution to the system. Consequently; applying traditional shapley value does not provide an attractive payoff to agents. In addition; fairness may also be lost because individual payoffs are less than singleton coalition values. We show how this environments may occur and how we can propose a more attractive and; still; fair payoffs to agents.

Today; the MEDLINE is an important repository containing more than 26 million citations and abstracts in the fields of medicine; while PubMed provides free access to MEDLINE and links to full-text articles. MEDLINE abstracts becomes a potential source of new knowledge in medical field. However; it is time-consuming and labour-intensive to find knowledge from MEDLINE abstracts; when a search returns much abstracts and each may contain a large volume of information. Therefore; this work aims to present a method of summarizing clinical knowledge from a MEDLINE abstract. The main mechanisms of the proposed method are driven on natural language processing (NLP) and text filtering techniques. The case study of this work is to summarize the clinical knowledge from a MEDLINE abstracts relating to cervical cancer in clinical trials. In the evaluation stage; the actual results obtained from a domain expert are used to compare the predicted results. After testing by recall; precision; and F-score; they return the satisfactory results; where the average of recall; precision; and F-measure are 0.84; 1.00; and 0.91 respectively.

We present a multiagent-based framework for crowd simulation in large space urban area on a standalone PC. We use Belief-Desire-Intention (BDI) for modeling individual agent behavior. We use RVO for handling a large number of agents. The simulation engine is Unity3d which also take care of the visualization. We experimented our framework with up to 20;000 agents; navigating them from origins to destinations. We found that we can navigate agents successfully. The execution time increases when the number of agent increase. The visualization becomes slow when the number of agent is higher than 1000 agents. We found that the the simulation steps also increases when the number of agent is not higher than 5005.